To Reduce Variability/Enhance Reproducibility: Begin at the Beginning

In May, my colleague, Jeanette Mowery, and I presented an InnovATEBIO webinar: To Reduce Variability/Enhance Reproducibility: Begin at the Beginning. This is the first of two blogs about that webinar.

Before discussing the webinar, I’ll share a personal story. Many years ago I was thrilled to land a postdoc position in a cancer research laboratory where I worked on a project involving the role of DNA adducts in initiating breast cancer. One of the reagents that I required had a pH of 10, and it included a relatively high concentration of urea. This reagent frustrated me. Every time I measured the pH of the solution, it had changed from the previous reading; I was never sure what the solution’s pH actually was.

Eventually I reached Alice, a technical support person employed by the pH meter manufacturer. I told her the recipe for my problematic reagent and Alice launched into a discussion of the special measuring electrode and the unusual reference electrode filling solution that my solution required. At some point I stopped her saying that I was looking at our pH meter and it only had one electrode, not two, as she was describing. I heard several seconds of silence on the line, and then she explained patiently in a sweet voice, suitable for a five year old, that, in fact, all pH meters used two electrodes. I was looking at a combination electrode in which two electrodes were inside one housing. She went on to explain that combination electrodes are suitable for some solutions, but not one with a high pH, nor one containing urea.

Based on Alice’s instructions, I obtained two separate electrodes and prepared and filled the reference electrode with the new filling solution. I then proceeded to accidentally tap the side of my beaker with the measuring electrode, thus shattering its fragile glass bulb and learning another important lesson about pH measurements. Fortunately, my next try with a new measuring electrode was successful, and the reagent’s measured pH was stable.

The point of this story, and the point of our webinar, is that avoiding variability and error requires mastering tasks that we often take for granted – like properly operating common laboratory equipment. Performing quality work requires beginning at the beginning. 

In our recent webinar, Jeanette and I presented quality control data that students obtained when they performed basic procedures, such as preparing salt solutions and buffers. We use conductivity as a quality control method to evaluate reagents. An example of student data is shown here:

Student quality control data for phosphate buffer, pH 7.2. This buffer is prepared by first separately making Solutions A and B and then combining the correct amounts of each to get a final buffer with the desired pH, in this case pH 7.2. The last two columns show the pH each team achieved for their final buffer solution and the conductivities of their final buffers. Observe that there is variability in both the pH values and the conductivities. Note also that the highest conductivity reading is more than twice that of the lowest reading.     

When students perform this and other laboratory exercises, they are often unaware that their products are not properly prepared until they see the class data on the blackboard. Exploring their data leads to discussions of fundamental questions, such as, for the phosphate buffer, why is there variability in their pH and conductivity values? Did some teams make mistakes and if so, what were they? How much variability in final conductivity values is acceptable? How can variability and error in reagent preparation be reduced? Students have rich conversations about fundamental laboratory practices – even when the tasks they performed are relatively simple.

Note that the students whose data are provided in the image above, and whose data were shown in the webinar, were in sections of our basic lab skills course reserved for individuals with at least a BS degree in biology. Many of these students have graduate degrees and/or industry experience. (Associate degree students take the same class but in separate sections.) These students’ quality control data provide a window into how variability and error creep into lab work, even when tasks are relatively simple, and even when the individuals involved are relatively experienced. The webinar contains more examples of student data from a variety of laboratory activities and the lessons that we can derive from them.

Stay tuned for a second blog that continues with a yet more entertaining (?) discussion of pH measurements.